2.1 BRD4 inhibitors and genetic silencing promoted ferroptosis induced by erastin in a panel of five cell lines, HEK293T, HeLa, HepG2, RKO, and PC3

First, we investigated the effect of BRD4 inhibitors on ferroptosis. Erastin is a ferroptosis inducer, and it induces ferroptosis in HEK293T and HeLa cells as evidenced by the cellular morphology and propidium iodide (PI) staining (Fig. 1A). JQ-1 and I-BET-762 are inhibitors of BRD4 [6]. Upon BRD4 inhibition by JQ-1 and I-BET-762, erastin-induced ferroptosis was greatly enhanced in HEK293T cells (Fig. 1B and Additional file 1: Figure S1A). We obtained similar results by using HeLa cells (Fig. 1B and Additional file 1: Figure S1A). To validate these results, we performed similar experiments on additional cell lines, HepG2 (hepatocellular carcinoma), RKO (colon carcinoma), and PC3 (prostatic adenocarcinoma). The results in these three cell lines were similar to those in 293 T and HeLa cells (Fig. 1C and Additional file 1: Figure S1B, S1C, and S1D). Erastin-induced cell death was greatly enhanced by JQ-1 or I-BET-762 treatment. To further confirm these results, we quantified the cell viability. As shown, erastin treatment reduced cell viability to 50%, which was significantly further reduced to 20% when co-treated with JQ-1 (Fig. 1D). Similar results were observed in these cell lines when another inhibitor, I-BET-762, was used, as well as in HeLa cells (Fig. 1D). To examine whether BRD4 played a critical role, we overexpressed BRD4-GFP in HEK293T cells. Compared to GFP-negative cells, GFP-positive cells were more resistant to erastin (Fig. 1E). These results demonstrated that BRD4 inhibition promoted ferroptosis.

Fig. 1

BRD4 alleviates ferroptosis induced by Erastin in a panel of five cell lines including HEK293T, HeLa, HepG2, RKO and PC3 cells. A Representative pictures and propidium iodide staining of HEK293T, HeLa, HepG2, RKO and PC3 cells after Erastin (20 μM) treatment for 24 h. Representative pictures and propidium iodide staining of (B) HEK293T cells, HeLa cells, (C) HepG2 cells, RKO cells and PC3 cells after being treated with DMSO, JQ-1 (1 μM), I-BET-762 (2 μM), Erastin (20 μM) or Erastin plus JQ-1, I-BET-762 respectively for 48 h. D CCK-8 was used to detect the cell viability of HEK293T cells and HeLa cells. E Representative pictures of HEK293T cells transfected with vector or plasmids encoding BRD4 after being treated with DMSO or Erastin (20 μM) respectively for 48 h. F Representative pictures of control and stable BRD4 knockdown cells in 293 T cells being treated with Erastin (20 μM) for 48 h. G Representative pictures of control and stable BRD4 knockdown cells in HeLa cells being treated with Erastin (20 μM) for 48 h. *p < 0 .05; **p < 0 .01; ***p < 0.001; ****p < 0.0001. Values are mean ± SD, n = 5

To exclude off-target effects of JQ-1 and I-BET-762, we generated stable BRD4 knockdown in 293 T and HeLa cells. Both RT-PCR and western blot confirmed the knockdown efficiency (Figure S2A and S2B). Then, we treated the control and stable BRD4 knockdown cells with erastin. As shown in Fig. 1F and 1G, erastin slightly induced cell death in control cells. However, in BRD4 knockdown cells, the erastin-induced cell death was greatly increased. Together, our results indicated that inhibition of BRD4 greatly promoted erastin-induced cell death.

2.2 Effect of BRD4 inhibitors on ferritin heavy chain 1 (FTH1) and ferritin light chain (FTL) in HEK293T and HeLa cells

The accumulation of excess free Fe2+ in cells is an important factor in ferroptosis. FTH1 and FTL promote the storage of Fe2+ in cells and maintain the homeostasis of iron [12, 13]. To understand the molecular mechanism of BRD4 inhibition, we examined key pathways associated with ferroptosis. We found that BRD4 inhibition upregulated the protein levels of FTH1 and FTL in HEK293T cells (Fig. 2A). Meanwhile, BRD4 inhibitors also activated the mRNA expression of FTH1 and FTL to varying degrees (Fig. 2B). In HeLa cells, both the protein and mRNA of FTH1 were downregulated after treatment with BRD4 inhibitors (Fig. 2C and 2D). However, FTL levels did not change significantly in HeLa cells. These studies suggest that BRD4 inhibitors downregulated FTH1 expression in certain cells. However, the reduced FTH1 expression upon BRD4 inhibition did not explain the common promoting effect on ferroptosis after BRD4 inhibition.

Fig. 2

Effect of BRD4 on FTH1 and FTL in HEK293T cells and HeLa cells. HEK293T cells and HeLa cells were treated with JQ-1 and I-BET-762 for 24 h. A The protein levels of FTH1 and FTL in HEK293T cells were detected by Western blot. B qRT-PCR analysis of the mRNA expression levels of FTH1 and FTL in HEK293T cells. C The protein levels of FTH1 and FTL in HeLa cells were detected by Western blot. D qRT-PCR analysis of the mRNA expression levels of FTH1 and FTL in HeLa cells. *p < 0 .05; **p < 0 .01; ***p < 0.001

2.3 BRD4 inhibitors or genetic silencing increased ROS and malondialdehyde levels of HEK293T and HeLa cells

ROS is a significant feature of ferroptosis in cells. Therefore, we continued to study whether the promotion of BRD4 inhibitors on ferroptosis of HEK293T and HeLa cells was related to ROS. We found that ROS increased significantly in HEK293T cells after treatment with JQ-1 (Fig. 3A). The ROS accumulation induced by JQ-1 was greater than that of the positive control, which was treated by Rosup. Another inhibitor, I-BET-762, showed a similar effect on ROS production (Fig. 3B). In HeLa cells, we also observed a similar effect on ROS production (Fig. 3C and 3D). To further validate the finding, we examined the ROS level in stable BRD4 knockdown cells. Compared to control cells, the stable BRD4 knockdown cells showed higher levels of ROS (Fig. 3E and 3F). Excess of ROS could attack biomembranes to induce lipid peroxidation. Thus, we examined the cellular level of malondialdehyde (MDA), which is a marker for lipid peroxidation and ferroptosis. Our results showed that JQ-1 or I-BET-762 treatment could significantly increase the level of cellular malondialdehyde (Fig. 3G). Furthermore, the level of MDA was also increased in BRD4 knocking down cells compared to control cells (Fig. 3H). These studies showed that BRD4 inhibitors induced the accumulation of ROS in both HEK293T and HeLa cells. The results suggest that BRD4 inhibitors may commonly trigger ROS and thus promote ferroptosis.

Fig. 3

Effect of BRD4 on ROS levels in HEK293T cells and HeLa cells. HEK293T cells and HeLa cells were treated with JQ-1 (3 μM) and I-BET-762 (5 μM) respectively for 24 h. Rosup (50 μg/mL) was used as positive control to treat cells for 1 h. Reactive oxygen species assay kit was used to assess the ROS level in cells. Representative pictures of ROS level in HEK293T cells after being treated with (A) JQ and B I-BET-762. Representative pictures of ROS level in HeLa cells after being treated with (C) JQ and D I-BET-762. E Representative pictures of ROS level in control and stable BRD4 knockdown cells in 293 T cells. F Representative pictures of ROS level in control and stable BRD4 knockdown cells in HeLa cells. G 293T and Hela cells were treated with JQ-1 or I-BET-762. The level of malondialdehyde (MDA) was measured in these cells. H The level of malondialdehyde (MDA) was measured in BRD4 knocking down cells

2.4 Effect of BRD4 inhibitors on VDAC2 and VDAC3 in HEK293T and HeLa cells

VDAC is an ion channel on the mitochondrial membrane that promotes ROS in mitochondria to enter the cytoplasm [4, 14]. The upregulation of ROS in cytoplasm mediated by VDAC2 and VDAC3 is one of the mechanisms by which erastin promotes ferroptosis in cells [4, 15]. However, we found that inhibition of BRD4 led to the downregulation of the expression of VDAC2 in HEK293T cells (Fig. 4A). At the same time, a BRD4 inhibitor also significantly reduced the mRNA levels of VDAC2 and VDAC3 (Fig. 4B). Similarly, the inhibition of BRD4 resulted in a downregulation of VDAC2 in HeLa cells (Fig. 4C). BRD4 inhibitors also significantly repressed the mRNA of VDAC2 and VDAC3 in HeLa cells (Fig. 4D). These experiments demonstrated that although BRD4 inhibitors significantly reduced the level of VDAC2 and VDAC3, they could not explain the greatly increased ROS production upon BRD4 inhibition.

Fig. 4

Effect of BRD4 on VDAC2 and VDAC3 in HEK293T cells and HeLa cells. HEK293T cells and HeLa cells were treated with JQ-1 and I-BET-762 for 24 h. A The protein level of VDAC2 in HEK293T cells was detected by Western blot. B qRT-PCR analysis of the mRNA expression levels of VDAC2 and VDAC3 in HEK293T cells. C The protein levels of VDAC2 in HeLa cells were detected by Western blot. D qRT-PCR analysis of the mRNA expression levels of VDAC2 and VDAC3 in HeLa cells. *p < 0 .05; **p < 0 .01; ***p < 0.001

2.5 Effect of BRD4 inhibitors on Nrf2 in HEK293T and HeLa cells

To further explore the specific mechanism by which BRD4 limited ROS production, we analyzed the expression of antioxidant Nrf2. Due to the presence of non-specific bands after incubation with the NRF2 antibody, we used the proteasome inhibitor MG132 to inhibit the ubiquitination degradation of NRF2 to determine the specific bands of NRF2 (Fig. 5A). We further examined the changes of NRF2 protein and mRNA levels in cells after treatment with BRD4 inhibitors JQ-1 and I-BET-762. The results showed that the Nrf2 protein levels in HEK293T cells increased significantly, but were significantly downregulated in HeLa cells (Fig. 5B). The changes in NRF2 mRNA levels in HEK293T and HeLa cells were similar to the changes in protein levels after treatment with JQ-1 or I-BET-762 (Fig. 5C and 5D). This suggests that BRD4 inhibitor-induced ROS production was not affected by NRF2.

Fig. 5

Effect of BRD4 on Nrf2 in HEK293T cells and HeLa cells. A Western blot was used to detect the protein level of Nrf2 in HEK293T cells and HeLa cells with MG132 treatment. HEK293T cells and HeLa cells were treated with JQ-1 and I-BET-762 for 24 h. BNFR2 protein levels were measured by Western blot. NRF2 mRNA levels in HEK293T cells (C) and HeLa cells (D) were determined by qRT-PCR. **p < 0 .01; ***p < 0.001

2.6 Effect of BRD4 inhibitors on system Xc and GPX4 in HEK293T and HeLa cells

GPX4 and system Xc, which includes SLC7A11 and SLC3A2 subunits, are important mechanisms for inhibiting ROS production and ferroptosis [1, 2, 16]. The protein level of GPX4 was slightly increased in HEK293T cells after JQ-1 treatment (Fig. 6A). The mRNA levels of SLC7A11, SLC3A2, and GPX4 in HEK293T cells were elevated after treatment with BRD4 inhibitors (Fig. 6B). However, the protein level of GPX4 in HeLa cells was not affected by BRD4 inhibitors (Fig. 6C). In HeLa cells, BRD4 inhibition significantly reduced the mRNA levels of SLC7A11 and SLC3A2, but BRD4 inhibition had minimal effect on the GPX4 mRNA level (Fig. 6D). This indicated that BRD4 inhibitors did not induce ROS production by decreasing the protein level of GPX4, SLC7A11, or SLC3A2 in either HEK293T or HeLa cells Additional file 2: Figure S2.

Fig. 6

Effect of BRD4 on System Xc and GPX4 in HEK293T cells and HeLa cells. HEK293T cells and HeLa cells were treated with JQ-1 and I-BET-762 for 24 h. A The protein level of GPX4 in HEK293T cells was detected by Western blot. B qRT-PCR analysis of the mRNA expression levels of SLC7A11, SLC3A2, and GPX4 in HEK293T cells. C Detected the protein levels of GPX4 in HeLa cells by Western blot. D The mRNA expression levels of SLC7A11, SLC3A2, and GPX4 in HeLa cells were measured by qRT-PCR. *p < 0 .05; **p < 0 .01; ***p < 0.001

2.7 BRD4 inhibitors decreased the level of FSP1 in HEK293T and HeLa cells

FSP1 is part of another GPX4-independent mechanism for inhibiting ROS production in ferroptosis [17]. Thus, we examined whether BRD4 inhibition affected FSP1 expression in cells. Importantly, JQ-1 treatment in HEK293T cells greatly reduced the protein level of FSP1 (Fig. 7A). Consistently, I-BET-762 also significantly reduced the protein level of FSP1 (Fig. 7A). Furthermore, the mRNA level of FSP1 was also greatly reduced upon JQ-1 and I-BET-762 treatment (Fig. 7B). Next, we determined the effect of BRD4 inhibition on FSP1 in HeLa cells. Consistent with the results from HEK293T cells, BRD4 inhibition greatly reduced the protein and mRNA level of FSP1 in HeLa cells (Fig. 7C and 7D). To further validate these results, we determined the level of FSP1 in control and stable BRD4 knockdown cells. The mRNA level of FSP1 was significantly reduced in stable BRD4 knockdown cells (Fig. 7E). Consistent with the downregulation of FSP1 mRNA, the protein level of FSP1 was also greatly reduced (Fig. 7F). To further confirm the direct regulation of BRD4 on FSP1, we analyzed the published BRD4 ChIP-sequencing data. In the FSP1 (AIFM2) promoter, we found significant enrichment of BRD4. Furthermore, the binding of BRD4 at the FSP1 promoter was greatly reduced (Fig. 7G). The results demonstrated a critical role of BRD4 on the expression of FSP1. To further confirm the effect of BRD4 inhibition on the exppression of FTH1, FTL, VDAC2, VDAC3, NRF2 and GPX4, we re-analyzed a large dataset from Connectivity Map, which tested ~ 5000 small-molecule compounds in 36 cell lines. Their tested compounds including I-BET-762 which is a compound used in our manuscript. We analyzed the expression of NFE2L2, FTH1, FTL, GPX4 and VDAC2 after treatment with I-BET-762 in 36 cell lines (Additional file 3: Figure S3). Consistent with our results, the expression of NFE2L2, FTH1, FTL, GPX4 and VDAC2 after treatment with I-BET-762 was inconsistent changed (Supplemental Figure S3A, S3B). In most cell lines, the expression of NFE2L2, FTH1, FTL, GPX4 and VDAC2 after treatment with I-BET-762 was up-regulated. By considering the consistent effect of BRD4 inhibition on FSP1, our findings suggest that BRD4 inhibition promotes ferroptosis very likely through FSP1.

Fig. 7

Effect of BRD4 on the expression of FSP1 in HEK293T cells and HeLa cells. HEK293T cells and HeLa cells were treated with JQ-1 and I-BET-762 for 24 h. The protein level of FSP1 in (A) HEK293T cells and (C) HeLa cells were detected by Western blot. The mRNA expression levels of FSP1 in (B) HEK293T cells and (D) HeLa cells were measured by qRT-PCR. E The mRNA levels of FSP1 in control and stable BRD4 knockdown cells in HEK293T and HeLa cells were measured by qRT-PCR. F The protein levels of FSP1 in control and stable BRD4 knockdown cells in HEK293T and HeLa cells were measured by Western blot. G The ChIP sequencing data of BRD4 in SUM-159 cells with (SRX2194233) or without JQ-1(SRX2194281) treatment was download from GEO dataset and the data was analyzed by using USCS genome. HeLa cells were measured by qRT-PCR.*p < 0 .05; **p < 0 .01; ***p < 0.001